Transactions on Machine Intelligence2821-16937320240401Motor Imagery EEG Signal Processing Using Common Spatial Patterns (CSP) and Python-Based Artificial Intelligence16116919414310.47176/TMI.2024.161ENM.LakzaeiFaculty Member, Chabahar Maritime University, Faculty of Marine Engineering, Department of Marine Electronics and Telecommunications, Chabahar, IranJournal Article20240307Brain–Computer Interface (BCI) systems create a direct communication channel between the human brain and external devices, bypassing conventional neuromuscular pathways. These systems interpret brain activity typically captured via electroencephalography (EEG) to infer user intent and execute commands accordingly. In this study, we focus on the classification of motor imagery (MI) signals, a widely used paradigm in BCI applications, which involves users imagining specific limb movements without actual muscle activation. EEG data corresponding to these imagined movements were preprocessed and analyzed using the Common Spatial Patterns (CSP) algorithm, a spatial filtering method that enhances class-discriminative features by maximizing variance differences across mental tasks. Subsequently, these features were classified using machine learning techniques implemented in the Python 3.7 environment. The EEG datasets used for training and evaluation were obtained from PhysioNet, a widely recognized repository hosted by the Massachusetts Institute of Technology (MIT). The aim of this work is to support the development of real-time, non-invasive BCI systems, with potential applications ranging from neurorehabilitation to the control of assistive devices such as prosthetics and exoskeletons. Additionally, the results offer insight into the implementation of neural signal processing algorithms on embedded systems, paving the way for the development of brain-controlled microchips and next-generation human–machine interfaces.https://www.tmachineintelligence.ir/article_194143_05c7c06682cc9e0d78588262d842f3de.pdfTransactions on Machine Intelligence2821-16937320240401Skin Melanoma Cancer Detection Using Particle Swarm Optimization Algorithm and Deep Learning17017819414410.47176/TMI.2024.170ENT.TorabiDepartment of Electrical Engineering, Hadaf Higher Education Institute, Sari, IranK.EsmaeiliDepartment of Electrical Engineering, K.N. Toosi University of Technology, Tehran, IranM.OmranDepartment of Electrical Engineering, Islamic Azad University, Sari Branch, Sari, IranS. M.AnishehAssistant Professor, Department of Electrical Engineering, Hadaf Higher Education Institute, Sari, IranJournal Article20230306Skin cancer is one of the most prevalent and life-threatening forms of cancer, with its incidence rapidly increasing over the past few decades. Early detection plays a crucial role in improving the survival rate of individuals diagnosed with skin cancer, particularly melanoma, which is the deadliest form. Traditionally, skin cancer diagnosis has relied on time-consuming and invasive methods such as skin biopsy. However, with advancements in technology, automated diagnosis through intelligent techniques has shown the potential to expedite the detection process and increase diagnostic accuracy. Among the various methods explored for skin cancer detection, Convolutional Neural Networks (CNNs) have emerged as one of the most effective deep learning models. CNNs are capable of learning and extracting intricate features from skin images, making them highly suitable for melanoma classification tasks. In this study, a deep CNN model is designed and evaluated for classifying melanoma from other skin lesions. Key parameters of the CNN, such as filter size and the number of filters, are optimized using the Particle Swarm Optimization (PSO) algorithm. This optimization process aims to minimize classification errors and enhance the overall performance of the model. The simulation results reveal that the proposed method outperforms existing frameworks, achieving a remarkable accuracy rate of 96% on the utilized dataset, demonstrating the effectiveness and reliability of the proposed approach for melanoma detection.https://www.tmachineintelligence.ir/article_194144_fd5dd579235334906f30b68f956a7314.pdfTransactions on Machine Intelligence2821-16937320240401Personal Recommender Model and Predicting Consumer Behavior in Digital Marketing Based on Deep Learning17919320510710.47176/TMI.2024.179ENA.MirzaeeDepartment of computer engineering, Afagh Institute of Higher Education, Urmia, IranM.ZeynaliDepartment of computer engineering, Islamic Azad University, Urmia Branch, Urmia, IranA.GhorbanzadehDepartment of computer engineering, Kamal Institute of Higher Education, Urmia, IranP.GhorbanzadehDepartment of computer engineering, Urmia University of Technology, Urmia, IranJournal Article20240330This study presents the design and evaluation of a personalized recommender system aimed at predicting consumer behavior within digital marketing environments. The proposed model integrates the strengths of Long Short-Term Memory (LSTM) networks and Recurrent Neural Networks (RNNs) to effectively process and learn from sequential data. LSTM units are employed to capture long-range temporal dependencies in user interactions, while RNN layers provide a framework for processing dynamic sequences of consumer activity over time. The model is trained using a dataset derived from real-world digital marketing platforms, which includes detailed logs of user preferences, purchase history, and browsing patterns. By learning from these behavioral indicators, the hybrid LSTM-RNN model is able to generate highly personalized recommendations tailored to individual consumers. Experimental results indicate that the proposed architecture achieves a high level of predictive accuracy, outperforming traditional recommendation methods in several key performance metrics such as precision, recall, and F1-score. These findings underscore the effectiveness of deep learning approaches in modeling complex consumer behaviors and highlight the potential of neural network-based recommender systems in optimizing marketing campaigns and enhancing user engagement. Ultimately, this research contributes to the advancement of intelligent recommendation technologies in the digital marketing domain, offering practical implications for businesses aiming to deliver more targeted and responsive customer experiences.https://www.tmachineintelligence.ir/article_205107_a6244149d2352e3092ad15e9654df042.pdfTransactions on Machine Intelligence2821-16937320240401Investigation of Creep Behavior of a Gas Turbine Bade with a Visco Plastic FEM Model to Estimate the Blade Life19420720510810.47176/TMI.2024.194ENM.ForghaniLecturer, School of engineering, Zand Institute of Higher education, Shiraz, Iran.K.AriafarAssistant Professor, School of engineering, Zand Institute of Higher education, Shiraz, Iran.Journal Article20240511Creep deformation is a major failure mechanism in turbine blades operating under high-temperature and high-stress conditions over extended periods. Traditional approaches to estimating the service life of turbine blades such as the Larson-Miller Parameter (LMP) method rely on simplified assumptions and offer only approximate predictions, which may not adequately reflect the complex time-dependent behavior of materials. In this study, a more accurate and physically realistic methodology is proposed using finite element (FE) analysis based on time-dependent plasticity to simulate creep in turbine blades. Three different constitutive models accounting for creep deformation are employed to evaluate their effectiveness in predicting the blade’s lifespan. The turbine blade is assumed to be composed of Inconel 738LC, a cast nickel-based superalloy widely used in aerospace and power generation applications due to its high-temperature strength and corrosion resistance. The simulation results are benchmarked against experimental creep data and compared to predictions obtained using the LMP method. The findings demonstrate that the proposed time-dependent plastic models provide a significantly improved prediction of creep life, showing a longer service duration compared to those estimated by conventional elastic stress analysis and the LMP approach. This modeling framework offers a more robust and accurate tool for the design and durability assessment of critical turbine components.https://www.tmachineintelligence.ir/article_205108_2b9034497b1480648e78fa8807cf0ddc.pdfTransactions on Machine Intelligence2821-16937320240401Increasing Error Detection in Software Testing Using Cuckoo Algorithm and Gravity Search Algorithm20822220510910.47176/TMI.2024.208ENS.MeydaniBehmenyar Institute of Higher Education, Kerman, IranF.SadeghiAssistant Professor, Bahmanyar Institute of Higher Education, Kerman, IranJournal Article20240526In the realm of software testing, one of the most critical challenges faced by development teams is the limitation of resources and time. As software systems grow in complexity and size, the number of test cases increases significantly, making it impractical to re-execute the entire suite of tests in each testing cycle. Consequently, there arises a need for effective strategies to select and prioritize test cases in a way that ensures the most valuable and error-prone parts of the software are tested early. Prioritizing test cases not only accelerates the detection of defects but also enhances the efficiency of the testing process by focusing efforts on areas with higher potential for failure. In this study, two powerful nature-inspired metaheuristic algorithms Cuckoo Search Optimization Algorithm and Gravitational Search Algorithm are employed to address the problem of test case prioritization. These algorithms are used to prioritize test cases based on coverage criteria, particularly aiming for maximum fault detection coverage. By optimizing the sequence of test execution, the proposed method improves both the fault detection rate and the overall effectiveness of the testing process. This approach contributes to the timely identification of critical defects and supports faster and more reliable software releases.https://www.tmachineintelligence.ir/article_205109_e776e48c1d0ca4f6e84ba99a23090a0e.pdfTransactions on Machine Intelligence2821-16937320240401An Outlier Detection Approach To Highlight Effective Genes By A Deep Learning Model and An Adjusted Genetic Algorithm (DLAGA)22323720511010.47176/TMI.2024.223ENY.AliakbarpoorDepartment of Computer Engineering, Shiraz Branch, Islamic Azad University, Shiraz, Iran0009-0008-3050-2796E.ParvinniaDepartment of Computer Engineering, Shiraz Branch, Islamic Azad University, Shiraz, Iran0000-0002-4728-5483Sh.SetayeshDepartment of Computer Engineering, Shiraz Branch, Islamic Azad University, Shiraz, Iran0009-0005-0628-5654Journal Article20240701Identifying abnormally expressed genes is a critical step in cancer diagnosis and has attracted significant attention within the biomedical research community. Gene expression datasets typically involve high-dimensional data, which poses major challenges during the pre-processing stage, particularly in maintaining the biological relevance and interpretability of selected genes. Traditional gene selection techniques often struggle with high computational demands and fail to preserve the intrinsic biological meaning of genes. In this study, we present an effective two-phase framework for gene selection and classification tailored for cancer diagnosis. The first phase employs a Variational Autoencoder (VAE), a deep learning-based technique, to reduce data dimensionality while capturing essential gene expression patterns. In the second phase, we utilize an Adjusted Genetic Algorithm (AGA) to search for a subset of informative genes. To further enhance classification performance, we integrate a wrapper-based approach within the AGA to individually classify genes relevant to different cancer types. Our method was evaluated on two publicly available microarray datasets. The experimental results reveal that the proposed framework outperforms several existing approaches in terms of classification accuracy, while maintaining reasonable computational efficiency. The integration of VAE and AGA offers a robust and biologically interpretable approach to gene selection, making it a promising tool for advancing precision oncology. These findings underscore the potential of combining deep learning and evolutionary algorithms for effective biomarker discovery in high-dimensional genomic data.https://www.tmachineintelligence.ir/article_205110_39c92b73e443fc21a2d75f4cae796133.pdf